An improved interface model for molten corium–concrete interaction Original Research Article

In the field of research for Severe Accidents of PWRs, calculation results are needed to estimate when corium achieves basemat melt-through. In this framework, knowledge of the heat transfer coefficient as well as the temperature at the interface between the melt and the solid are key issues. It has been previously emphasized that physico-chemistry of the melt (composition) affects the temperature. However, the effect of gas sparging and liquid concrete release on the mechanical stability of the solid layer and on the interface temperature has not been analysed in detail. ARTEMIS 1D tests were launched to analyse the phenomenology at the interface. A general conclusion for all these tests is that a solid medium enriched in refractory species exists at the interface between the pool and the concrete. Tests 2 and 6, which are close to reactor representative conditions, could be well described with the assumption Tinterface = Tliquidus. The analysis of tests 3 and 4 revealed that the interfacial medium is thicker by a factor of 4–5 times than calculated with the assumption of pure conduction heat transfer. Detailed analysis leads to the conclusion that the interfacial medium is made of a porous layer of refractory particles embedded in a liquid phase. The liquid density difference, between the porous medium and chimneys, results in recirculation of the liquid in the porous layer. This recirculation is responsible for a convective contribution to heat transfer through the porous layer. The convective heat flux is partly linked to cooling of the recirculating liquid, but also to its partial solidification. The solidification results in gradual plugging and enrichment of the porous layer in refractory species and in the increase of the resistance to heat transfer. The phenomena have been modelled and ARTEMIS tests 3 and 4 are well reproduced. The complete plugging of the porous medium, associated with thermodynamic equilibrium at the interface between the solid medium and the pool leads naturally to the solid crust model implemented in TOLBIAC-ICB. The latter model appears, thus, as a simplification of the porous medium approach described here. Thick interfacial layers, which have composition between refractory and corium mixture, have been observed in ACE and MACE tests with LCS (Limestone–Common Sand) or Limestone concrete and can be explained with the proposed model approach. For siliceous concrete, convection within a porous medium is not possible due to the large viscosity.